Cloud platforms have grown significantly in the last 10 years and are expected to continue to do so for the next couple decades. Businesses of all sizes have been investing in these cloud platforms.  

However, the promised cost savings continue to be elusive for many enterprises, which are faced with surprises like orders of magnitude increases in cloud-computing costs, increased cybersecurity exposures and system performance/outage issues. Such incidents are making CEOs second guess the cloud-adoption strategy.

The promise of these cloud platforms is real.  But I believe the adverse experiences many businesses are dealing with are the result of a not-well-thought-through approach, the absence of a clear business case and the lack of the right skill set by the team who is migrating systems and data to the cloud.  

Common Cloud-Adoption Issues

There are three major issues that can occur when a company attempts to transition to the cloud without taking the proper precautions. These include:

1. Questionable return on investment.
2. Higher computing costs.
3. System downtime.

1. Questionable Return on Investment

Cloud adoption projects get funded primarily on the premise that IT infrastructure costs will go down by moving to the cloud. In many cases, this cost reduction is not happening. According to a survey by McKinsey, 75 percent of the cloud migration costs are over budget and 38 percent of the projects are behind schedule. This casts serious doubt on future cloud adoption projects.

2. Higher Computing Cost 

One of the big promises of the cloud is pay-per-use. If a retailer needs peak capacity servers during the holiday shopping season, they can pay for peak capacity just for that time of year. While this opportunity can be seized on paper, there are instances where such variable pricing is causing unexpected order-of-magnitude increases in cloud bills due to incidents like a midnight job going in a circular loop and causing the computing need to skyrocket overnight.  One bad incident can put a brake on the cloud adoption movement. 

3.  Systems Downtime 

Computing environment and configuration parameters are similar between on-premise infrastructure and cloud infrastructure.  However, they are not the same.  When on-premise administrators migrate to the cloud without prior practical knowledge, it’s common for them to miss many items. While these items may be subtle, they often cause the application to stop running and become unavailable causing downtimes. 

More on Cloud ComputingWhy All Developers Want Their IT Infrastructure to Adopt Kubernetes

3 Tips for Improving your Cloud-Adoption Strategy

Before adopting the cloud, there are two things every business needs to do.  The first is to build a business case before they start any cloud adoption project. Bring experienced resources to plan and execute the project. The second is to upgrade the skills of existing staff to take over ongoing maintenance and upkeep of the cloud platform.

Here are three suggestions for improving your cloud adoption:

 

1. Put the Business Plan First

More often than not, cloud adoption projects are initiated by the IT team. As a result, cloud adoption can sometimes become a “solution looking for a problem.”  

Building a business case needs to be the core theme of any cloud adoption strategy. Building a business case should be focused on identifying tangible, quantifiable value generation and avoiding too much focus on the intangible value. Tangible value generation needs to be quantified in terms of revenue increase, reduction in cost and reduction in risks.  Any other value creation should be avoided.

Maintaining objectivity to carefully analyze different cloud options needs to be a critical step in the process. Not every situation is conducive to cloud adoption. There are few instances where adopting the cloud may not make sense. If the users don’t have reliable internet, cloud adoption doesn’t make sense. Highly sensitive data is another reason to be skeptical about cloud adoption. If the customer has legacy applications which are still on mainframe or AS2 machines, cloud adoption may not be an easy answer.

It’s important to walk away from such a project if the return on investment is not clear or if the risks inherent are high.

2. Understand the Cloud Platform’s Revenue Model  

Cloud is architected on the core principle of “pay-per-use.”  Pay-per-use is a cost-compelling idea and can be a win-win for the customers and for the cloud providers. 

For example, in a traditional data center, a retailer is expected to have a server-computing capacity equal to the highest demand of the year. With cloud computing, the same retailer could pay for a lower computing cost during off-peak season and throttle the computing power up during holiday seasons, driving significant cost savings.  

While the promise of the cloud is true, a simple “lift and shift” approach to moving your applications can bring a few nasty surprises. For example, before migration, one of our customer teams expected a $50,000 a month Azure bill. The actual bill for the first month turned out to be $180,000. The biggest reason for such a fluctuation was that the team decided to replicate the on-premise environments on the cloud. The team didn’t put up some basic guardrails. As a result, rogue back end jobs kept running infinite loops and the meter kept running. 

The migration team didn’t realize that they used to own the car and now they are leasing the car. If they keep running the car in infinite loops, they have to pay a penalty for the usage. 

The migration team should take into account the architectural and configuration differences between an on-premise server and a cloud server. Understanding the revenue model the cloud company has in place will help you optimize resource use and cost. Lift and shift may be easy, but it can still prove to be costly.

More on Cloud ComputingWhat Is Cloud Architecture? Understanding the Fundamentals.

3. Reskill Your IT Staff for the Cloud

Just because your IT infrastructure team worked well for you in the past does not mean that their skill set will help them be equally efficient in a cloud environment. A reskilling program needs to be in place to mitigate risks around cybersecurity, cost optimization, cloud configuration, and cloud architecture. 

Cloud vendors have published both entry level and expert level certifications. Every person working on cloud adoption should achieve vendor specific accreditation, first at the entry then at the expert level. 

After achieving those certifications, they need to go through practical use in a sandbox environment before trying to migrate real applications into the cloud platform. It’s always a good idea to bring in expert service providers for the first go around or two so that internal staff will have some time to finish accreditation and learn from the experts. While cloud vendors do offer help, it’s typically not enough for the internal team to feel confident without training.

In most cases, companies are taking shortcuts to upskill their staff, resulting in cost overruns, security exposures and system outages.  

As business owners or CEOs review these concerns, it’s natural that they are tempted to reverse course, saying farewell to the cloud and returning to their own servers for data storage.

However, reversing course is not an optimal decision most of the time. Staying in your own data centers is often suboptimal from both a cost and risk standpoint. To make the most of the cloud, businesses need to make the right decisions that will help them reduce expenses, address cybersecurity issues and make sure their technology teams are brought up to speed on the skills they need to do their jobs.